The Effect of Global and Local Attention Bias on Low Prevalence Search
Abstract
Fields that require visual search for rare target items are particularly vulnerable to the low-prevalence effect (LPE). This means that TSA officers and radiologists can often miss the target they are looking for because of an induced bias that the target won’t be found. The LPE is notoriously difficult to mitigate. The current experiment attempted to implicitly improve performance in low-prevalence conditions by training participants to adopt a global or local attentional bias, via the Navon task. We hypothesized that the local processing bias would increase the number of items inspected and consequently increase search times and performance. In experiment 1 (N = 45), participants completed the Navon task and the search task in separate blocks. The manipulation did not yield an effect on the LPE; however, a trend emerged that encouraged us to investigate further. In experiment 2 (N = 51), participants completed a randomized mixture of the Navon task and the search task. This experimental design was able to slightly increase the number of items inspected in the visual search when participants adopted the local processing bias compared to the global processing bias; however, the manipulation did not extend to search performance. Nonetheless, these findings suggest that implicit manipulations may be able to impact search behavior in low-prevalence conditions. Future studies should continue to investigate implicit forms of biasing and training to induce more robust effects on the LPE.
Related
- textsc{Athanor}: local search over abstract constraint specifications
- Small-LCM Distinct Covering Systems via Segmented Local Search
- Local Manifold Attention for Recovery and Generation of Rare 3D Human Poses
- Local Search for Almost-Spanning Square Grids in Erdős--Rényi Random Graphs
- GPU Parallel Pareto Local Search for Multi-objective QUBO
BibTeX
@misc{kelly2026effect,
title = {{The Effect of Global and Local Attention Bias on Low Prevalence Search}},
author = {Charlotte M Kelly and Robin Celani and Juan Guevara Pinto},
howpublished = {PsyArXiv},
year = {2026},
month = sep,
doi = {10.31234/osf.io/ydwzb_v2},
url = {https://osf.io/preprints/psyarxiv/ydwzb_v2/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - The Effect of Global and Local Attention Bias on Low Prevalence Search
AU - Charlotte M Kelly
AU - Robin Celani
AU - Juan Guevara Pinto
PY - 2026
DA - 2026/09/30
PB - PsyArXiv
DO - 10.31234/osf.io/ydwzb_v2
UR - https://osf.io/preprints/psyarxiv/ydwzb_v2/
AB - Fields that require visual search for rare target items are particularly vulnerable to the low-prevalence effect (LPE). This means that TSA officers and radiologists can often miss the target they are looking for because of an induced bias that the target won’t be found. The LPE is notoriously difficult to mitigate. The current experiment attempted to implicitly improve performance in low-prevalence conditions by training participants to adopt a global or local attentional bias, via the Navon task. We hypothesized that the local processing bias would increase the number of items inspected and consequently increase search times and performance. In experiment 1 (N = 45), participants completed the Navon task and the search task in separate blocks. The manipulation did not yield an effect on the LPE; however, a trend emerged that encouraged us to investigate further. In experiment 2 (N = 51), participants completed a randomized mixture of the Navon task and the search task. This experimental design was able to slightly increase the number of items inspected in the visual search when participants adopted the local processing bias compared to the global processing bias; however, the manipulation did not extend to search performance. Nonetheless, these findings suggest that implicit manipulations may be able to impact search behavior in low-prevalence conditions. Future studies should continue to investigate implicit forms of biasing and training to induce more robust effects on the LPE.
ER -